dnddnjs / pytorch-cifar10

The state-of-the-art algorithms on CIFAR-10 dataset in PyTorch

pytorch-cifar10

State of the art algorithm codes on CIFAR-10 dataset in PyTorch

State of the art models

• ResNet: paper, post, trained model download(92.5% accuracy)

  • network architecture
  • train & test error rate

• PyramidNet: paper, post, trained model download(95.19% accuracy)

  • network architecture
  • train & test error rate

• Shake-Drop: paper, post

• ENAS: paper, post